Message ID | 1583807817-5571-1-git-send-email-xiubli@redhat.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | ceph: add min/max latency support for read/write/metadata metrics | expand |
On Mon, 2020-03-09 at 22:36 -0400, xiubli@redhat.com wrote: > From: Xiubo Li <xiubli@redhat.com> > > These will be very useful help diagnose problems. > > URL: https://tracker.ceph.com/issues/44533 > Signed-off-by: Xiubo Li <xiubli@redhat.com> > --- > > The output will be like: > > # cat /sys/kernel/debug/ceph/19e31430-fc65-4aa1-99cf-2c8eaaafd451.client4347/metrics > item total sum_lat(us) avg_lat(us) min_lat(us) max_lat(us) > ------------------------------------------------------------------------------------- > read 27 297000 11000 2000 27000 > write 16 3860000 241250 175000 263000 > metadata 3 30000 10000 2000 16000 > > item total miss hit > ------------------------------------------------- > d_lease 2 0 1 > caps 2 0 3078 > > > > fs/ceph/debugfs.c | 27 ++++++++++++++++++++------ > fs/ceph/mds_client.c | 12 ++++++++++++ > fs/ceph/metric.h | 54 +++++++++++++++++++++++++++++++++++++++++++++++++++- > 3 files changed, 86 insertions(+), 7 deletions(-) > > > diff --git a/fs/ceph/metric.h b/fs/ceph/metric.h > index faba142..9f0d050 100644 > --- a/fs/ceph/metric.h > +++ b/fs/ceph/metric.h > @@ -2,6 +2,10 @@ > #ifndef _FS_CEPH_MDS_METRIC_H > #define _FS_CEPH_MDS_METRIC_H > > +#include <linux/atomic.h> > +#include <linux/percpu.h> > +#include <linux/spinlock.h> > + > /* This is the global metrics */ > struct ceph_client_metric { > atomic64_t total_dentries; > @@ -13,12 +17,21 @@ struct ceph_client_metric { > > struct percpu_counter total_reads; > struct percpu_counter read_latency_sum; > + spinlock_t read_latency_lock; > + atomic64_t read_latency_min; > + atomic64_t read_latency_max; > > struct percpu_counter total_writes; > struct percpu_counter write_latency_sum; > + spinlock_t write_latency_lock; > + atomic64_t write_latency_min; > + atomic64_t write_latency_max; > > struct percpu_counter total_metadatas; > struct percpu_counter metadata_latency_sum; > + spinlock_t metadata_latency_lock; > + atomic64_t metadata_latency_min; > + atomic64_t metadata_latency_max; > }; > > static inline void ceph_update_cap_hit(struct ceph_client_metric *m) > @@ -36,11 +49,24 @@ static inline void ceph_update_read_latency(struct ceph_client_metric *m, > unsigned long r_end, > int rc) > { > + unsigned long lat = r_end - r_start; > + > if (rc < 0 && rc != -ENOENT && rc != -ETIMEDOUT) > return; > > percpu_counter_inc(&m->total_reads); > - percpu_counter_add(&m->read_latency_sum, r_end - r_start); > + percpu_counter_add(&m->read_latency_sum, lat); > + > + if (lat >= atomic64_read(&m->read_latency_min) && > + lat <= atomic64_read(&m->read_latency_max)) > + return; > + > + spin_lock(&m->read_latency_lock); > + if (lat < atomic64_read(&m->read_latency_min)) > + atomic64_set(&m->read_latency_min, lat); > + if (lat > atomic64_read(&m->read_latency_max)) > + atomic64_set(&m->read_latency_max, lat); > + spin_unlock(&m->read_latency_lock); > } > Looks reasonable overall. I do sort of wonder if we really need spinlocks for these though. Might it be more efficient to use cmpxchg instead? i.e.: cur = atomic64_read(&m->read_latency_min); do { old = cur; if (likely(lat >= old)) break; } while ((cur = atomic_long_cmpxchg(&m->read_latency_min, old, lat)) != old); ...another idea might be to use a seqlock and non-atomic vars. Mostly this shouldn't matter much though as we'll almost always be hitting the non-locking fastpath. I'll plan to merge this as-is unless you want to rework it.
On 2020/3/16 22:21, Jeff Layton wrote: > On Mon, 2020-03-09 at 22:36 -0400, xiubli@redhat.com wrote: >> From: Xiubo Li <xiubli@redhat.com> >> >> These will be very useful help diagnose problems. >> >> URL: https://tracker.ceph.com/issues/44533 >> Signed-off-by: Xiubo Li <xiubli@redhat.com> >> --- >> >> The output will be like: >> >> # cat /sys/kernel/debug/ceph/19e31430-fc65-4aa1-99cf-2c8eaaafd451.client4347/metrics >> item total sum_lat(us) avg_lat(us) min_lat(us) max_lat(us) >> ------------------------------------------------------------------------------------- >> read 27 297000 11000 2000 27000 >> write 16 3860000 241250 175000 263000 >> metadata 3 30000 10000 2000 16000 >> >> item total miss hit >> ------------------------------------------------- >> d_lease 2 0 1 >> caps 2 0 3078 >> >> >> >> fs/ceph/debugfs.c | 27 ++++++++++++++++++++------ >> fs/ceph/mds_client.c | 12 ++++++++++++ >> fs/ceph/metric.h | 54 +++++++++++++++++++++++++++++++++++++++++++++++++++- >> 3 files changed, 86 insertions(+), 7 deletions(-) >> >> >> diff --git a/fs/ceph/metric.h b/fs/ceph/metric.h >> index faba142..9f0d050 100644 >> --- a/fs/ceph/metric.h >> +++ b/fs/ceph/metric.h >> @@ -2,6 +2,10 @@ >> #ifndef _FS_CEPH_MDS_METRIC_H >> #define _FS_CEPH_MDS_METRIC_H >> >> +#include <linux/atomic.h> >> +#include <linux/percpu.h> >> +#include <linux/spinlock.h> >> + >> /* This is the global metrics */ >> struct ceph_client_metric { >> atomic64_t total_dentries; >> @@ -13,12 +17,21 @@ struct ceph_client_metric { >> >> struct percpu_counter total_reads; >> struct percpu_counter read_latency_sum; >> + spinlock_t read_latency_lock; >> + atomic64_t read_latency_min; >> + atomic64_t read_latency_max; >> >> struct percpu_counter total_writes; >> struct percpu_counter write_latency_sum; >> + spinlock_t write_latency_lock; >> + atomic64_t write_latency_min; >> + atomic64_t write_latency_max; >> >> struct percpu_counter total_metadatas; >> struct percpu_counter metadata_latency_sum; >> + spinlock_t metadata_latency_lock; >> + atomic64_t metadata_latency_min; >> + atomic64_t metadata_latency_max; >> }; >> >> static inline void ceph_update_cap_hit(struct ceph_client_metric *m) >> @@ -36,11 +49,24 @@ static inline void ceph_update_read_latency(struct ceph_client_metric *m, >> unsigned long r_end, >> int rc) >> { >> + unsigned long lat = r_end - r_start; >> + >> if (rc < 0 && rc != -ENOENT && rc != -ETIMEDOUT) >> return; >> >> percpu_counter_inc(&m->total_reads); >> - percpu_counter_add(&m->read_latency_sum, r_end - r_start); >> + percpu_counter_add(&m->read_latency_sum, lat); >> + >> + if (lat >= atomic64_read(&m->read_latency_min) && >> + lat <= atomic64_read(&m->read_latency_max)) >> + return; >> + >> + spin_lock(&m->read_latency_lock); >> + if (lat < atomic64_read(&m->read_latency_min)) >> + atomic64_set(&m->read_latency_min, lat); >> + if (lat > atomic64_read(&m->read_latency_max)) >> + atomic64_set(&m->read_latency_max, lat); >> + spin_unlock(&m->read_latency_lock); >> } >> > Looks reasonable overall. I do sort of wonder if we really need > spinlocks for these though. Might it be more efficient to use cmpxchg > instead? i.e.: > > cur = atomic64_read(&m->read_latency_min); > do { > old = cur; > if (likely(lat >= old)) > break; > } while ((cur = atomic_long_cmpxchg(&m->read_latency_min, old, lat)) != old); IMO the above case should be more efficient here. Let me post the V2 to fix it. BRs > ...another idea might be to use a seqlock and non-atomic vars. > > Mostly this shouldn't matter much though as we'll almost always be > hitting the non-locking fastpath. I'll plan to merge this as-is unless > you want to rework it.
diff --git a/fs/ceph/debugfs.c b/fs/ceph/debugfs.c index 60f3e307..9ef0ffe 100644 --- a/fs/ceph/debugfs.c +++ b/fs/ceph/debugfs.c @@ -129,28 +129,43 @@ static int metric_show(struct seq_file *s, void *p) struct ceph_fs_client *fsc = s->private; struct ceph_mds_client *mdsc = fsc->mdsc; int i, nr_caps = 0; - s64 total, sum, avg = 0; + s64 total, sum, avg = 0, min, max; - seq_printf(s, "item total sum_lat(us) avg_lat(us)\n"); - seq_printf(s, "-----------------------------------------------------\n"); + seq_printf(s, "item total sum_lat(us) avg_lat(us) min_lat(us) max_lat(us)\n"); + seq_printf(s, "-------------------------------------------------------------------------------------\n"); total = percpu_counter_sum(&mdsc->metric.total_reads); sum = percpu_counter_sum(&mdsc->metric.read_latency_sum); sum = jiffies_to_usecs(sum); avg = total ? sum / total : 0; - seq_printf(s, "%-14s%-12lld%-16lld%lld\n", "read", total, sum, avg); + min = atomic64_read(&mdsc->metric.read_latency_min); + min = jiffies_to_usecs(min == S64_MAX ? 0 : min); + max = atomic64_read(&mdsc->metric.read_latency_max); + max = jiffies_to_usecs(max); + seq_printf(s, "%-14s%-12lld%-16lld%-16lld%-16lld%lld\n", "read", + total, sum, avg, min, max); total = percpu_counter_sum(&mdsc->metric.total_writes); sum = percpu_counter_sum(&mdsc->metric.write_latency_sum); sum = jiffies_to_usecs(sum); avg = total ? sum / total : 0; - seq_printf(s, "%-14s%-12lld%-16lld%lld\n", "write", total, sum, avg); + min = atomic64_read(&mdsc->metric.write_latency_min); + min = jiffies_to_usecs(min == S64_MAX ? 0 : min); + max = atomic64_read(&mdsc->metric.write_latency_max); + max = jiffies_to_usecs(max); + seq_printf(s, "%-14s%-12lld%-16lld%-16lld%-16lld%lld\n", "write", + total, sum, avg, min, max); total = percpu_counter_sum(&mdsc->metric.total_metadatas); sum = percpu_counter_sum(&mdsc->metric.metadata_latency_sum); sum = jiffies_to_usecs(sum); avg = total ? sum / total : 0; - seq_printf(s, "%-14s%-12lld%-16lld%lld\n", "metadata", total, sum, avg); + min = atomic64_read(&mdsc->metric.metadata_latency_min); + min = jiffies_to_usecs(min == S64_MAX ? 0 : min); + max = atomic64_read(&mdsc->metric.metadata_latency_max); + max = jiffies_to_usecs(max); + seq_printf(s, "%-14s%-12lld%-16lld%-16lld%-16lld%lld\n", "metadata", + total, sum, avg, min, max); seq_printf(s, "\n"); seq_printf(s, "item total miss hit\n"); diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c index 5c03ed3..ff6c2be 100644 --- a/fs/ceph/mds_client.c +++ b/fs/ceph/mds_client.c @@ -4358,6 +4358,10 @@ static int ceph_mdsc_metric_init(struct ceph_client_metric *metric) if (ret) goto err_read_latency_sum; + spin_lock_init(&metric->read_latency_lock); + atomic64_set(&metric->read_latency_min, S64_MAX); + atomic64_set(&metric->read_latency_max, 0); + ret = percpu_counter_init(&metric->total_writes, 0, GFP_KERNEL); if (ret) goto err_total_writes; @@ -4366,6 +4370,10 @@ static int ceph_mdsc_metric_init(struct ceph_client_metric *metric) if (ret) goto err_write_latency_sum; + spin_lock_init(&metric->write_latency_lock); + atomic64_set(&metric->write_latency_min, S64_MAX); + atomic64_set(&metric->write_latency_max, 0); + ret = percpu_counter_init(&metric->total_metadatas, 0, GFP_KERNEL); if (ret) goto err_total_metadatas; @@ -4374,6 +4382,10 @@ static int ceph_mdsc_metric_init(struct ceph_client_metric *metric) if (ret) goto err_metadata_latency_sum; + spin_lock_init(&metric->metadata_latency_lock); + atomic64_set(&metric->metadata_latency_min, S64_MAX); + atomic64_set(&metric->metadata_latency_max, 0); + return 0; err_metadata_latency_sum: diff --git a/fs/ceph/metric.h b/fs/ceph/metric.h index faba142..9f0d050 100644 --- a/fs/ceph/metric.h +++ b/fs/ceph/metric.h @@ -2,6 +2,10 @@ #ifndef _FS_CEPH_MDS_METRIC_H #define _FS_CEPH_MDS_METRIC_H +#include <linux/atomic.h> +#include <linux/percpu.h> +#include <linux/spinlock.h> + /* This is the global metrics */ struct ceph_client_metric { atomic64_t total_dentries; @@ -13,12 +17,21 @@ struct ceph_client_metric { struct percpu_counter total_reads; struct percpu_counter read_latency_sum; + spinlock_t read_latency_lock; + atomic64_t read_latency_min; + atomic64_t read_latency_max; struct percpu_counter total_writes; struct percpu_counter write_latency_sum; + spinlock_t write_latency_lock; + atomic64_t write_latency_min; + atomic64_t write_latency_max; struct percpu_counter total_metadatas; struct percpu_counter metadata_latency_sum; + spinlock_t metadata_latency_lock; + atomic64_t metadata_latency_min; + atomic64_t metadata_latency_max; }; static inline void ceph_update_cap_hit(struct ceph_client_metric *m) @@ -36,11 +49,24 @@ static inline void ceph_update_read_latency(struct ceph_client_metric *m, unsigned long r_end, int rc) { + unsigned long lat = r_end - r_start; + if (rc < 0 && rc != -ENOENT && rc != -ETIMEDOUT) return; percpu_counter_inc(&m->total_reads); - percpu_counter_add(&m->read_latency_sum, r_end - r_start); + percpu_counter_add(&m->read_latency_sum, lat); + + if (lat >= atomic64_read(&m->read_latency_min) && + lat <= atomic64_read(&m->read_latency_max)) + return; + + spin_lock(&m->read_latency_lock); + if (lat < atomic64_read(&m->read_latency_min)) + atomic64_set(&m->read_latency_min, lat); + if (lat > atomic64_read(&m->read_latency_max)) + atomic64_set(&m->read_latency_max, lat); + spin_unlock(&m->read_latency_lock); } static inline void ceph_update_write_latency(struct ceph_client_metric *m, @@ -48,11 +74,24 @@ static inline void ceph_update_write_latency(struct ceph_client_metric *m, unsigned long r_end, int rc) { + unsigned long lat = r_end - r_start; + if (rc && rc != -ETIMEDOUT) return; percpu_counter_inc(&m->total_writes); percpu_counter_add(&m->write_latency_sum, r_end - r_start); + + if (lat >= atomic64_read(&m->write_latency_min) && + lat <= atomic64_read(&m->write_latency_max)) + return; + + spin_lock(&m->write_latency_lock); + if (lat < atomic64_read(&m->write_latency_min)) + atomic64_set(&m->write_latency_min, lat); + if (lat > atomic64_read(&m->write_latency_max)) + atomic64_set(&m->write_latency_max, lat); + spin_unlock(&m->write_latency_lock); } static inline void ceph_update_metadata_latency(struct ceph_client_metric *m, @@ -60,10 +99,23 @@ static inline void ceph_update_metadata_latency(struct ceph_client_metric *m, unsigned long r_end, int rc) { + unsigned long lat = r_end - r_start; + if (rc && rc != -ENOENT) return; percpu_counter_inc(&m->total_metadatas); percpu_counter_add(&m->metadata_latency_sum, r_end - r_start); + + if (lat >= atomic64_read(&m->metadata_latency_min) && + lat <= atomic64_read(&m->metadata_latency_max)) + return; + + spin_lock(&m->metadata_latency_lock); + if (lat < atomic64_read(&m->metadata_latency_min)) + atomic64_set(&m->metadata_latency_min, lat); + if (lat > atomic64_read(&m->metadata_latency_max)) + atomic64_set(&m->metadata_latency_max, lat); + spin_unlock(&m->metadata_latency_lock); } #endif /* _FS_CEPH_MDS_METRIC_H */